Mapping CAC susceptibility loci to atherosclerotic plaques

Project ID: AE_20200512_COL_MKAVOUSI_MBOS_CHARGE_1000G_CAC.

Background

Collaboration for mapping targets from the CHARGE Consortium 1000G GWAS on coronary artery calcification (CAC) with Maryam Kavousi, Patricia Peyser, and Maxime Bos.

Athero-Express Biobank Study

We have bulk RNAseq (n = 635 samples) and single-cell RNAseq data, genome-wide methylation (Illumina 450K) in n ± 600, as well as overlapping genetic data for ±2,100 individuals with extensive histological plaque characterisation. We will address the following questions:

  • Gene expression correlated to characteristics of plaques?
  • Where target genes expressed, which cell types?
  • Are any of the variants associated to plaque characteristics?

Ideally, we would like to map the variants to the expression and methylation data (molecular QTL mapping), but this is pending finalizing our molQTL mapping.

We expect new plasma-based OLINK data (for the ‘TO_AITION’ projecty) in Q4 2020. We agreed that we would keep this in mind, we could run some additional analyses when asked during the revision stage of the manuscript.

  • Correlation of variants to proteins from the OLINK-platform

Methods

Mapping genes to single cells

We will use the last dataset from the scRNAseq data, including 35 individuals, to project target genes.

Genetic analyses

For the genetic analyses we will perform regression analyses adjusted for age, sex (where applicable) and principal components. So, we will apply the following model:

model 1: phenotype ~ age + sex + chip-used + PC1 + PC2 + year-of-surgery

phenotypes are:

  • calcification, coded Calc.bin no/minor vs. moderate/heavy staining
  • collagen, coded Collagen.bin no/minor vs. moderate/heavy staining
  • fat10, coded Fat.bin_10 no/<10% fat vs. >10% fat
  • fat40, coded Fat.bin_40 no/<40% fat vs. >40% fat
  • intraplaque hemorrhage, coded IPH.bin no vs. yes
  • macrophages (CD68), coded macmean0 mean of computer-assisted calculation CD68+ region of interest
  • smooth muscle cells (alpha-actin), coded smcmean0 mean of computer-assisted calculation SMA+ region of interest
  • intraplaque vessel density (CD34), coded vessel_density manually counted CD34+ cells per 3-4 hotspots
  • mast cells, coded Mast_cells_plaque manually counted mast cell tryptase+ cells (https://academic.oup.com/eurheartj/article/34/48/3699/484981)
  • neutrophils (CD66b), coded neutrophils manually counted CD66b+ cells (https://pubmed.ncbi.nlm.nih.gov/20595650/)

Continuous variables were inverse-rank normal transformated, indicated by _rankNorm.

Gene expression

Whole-plaque RNAseq

For the expression analysis we used carotid plaque-derived bulk RNAseq data and queried it for the gene list. Below a graph showing the overall expression of the genes (not all are in the data) compared to the mean expression of 1,000 randomly picked genes.

Figure 1: Overall expression of target genes in carotid plaques from the Athero-Express Biobank Study Overall expression of target genes in carotid plaques from the Athero-Express Biobank Study

We assessed the correlation with plaque characteristics (mentioned above) and secondary major adverse cardiovascular events (MACE [major]) at 30 days and 3 years after CEA.

Single cell RNAseq

We projected target genes to the single-cell RNAseq data derived from 37 carotid plaque samples. We identified cell communities (Figure 2), mapped and projected target gene expression to the cell communities (Figure 3).

Figure 2: Cell communities identified in carotid plaques from the Athero-Express Biobank Study Cell communities identified in carotid plaques from the Athero-Express Biobank Study

Figure 3: Dotplot showing expression of target genes per cell type in carotid plaques from the Athero-Express Biobank Study Dotplot showing expression of target genes per cell type in carotid plaques from the Athero-Express Biobank Study

QTL mapping

Quantitative trait loci mapping has not finished. This could be added at a later stage.

The MIT License (MIT)